Predicting type of delivery by identification of obstetric risk factors through data mining

In Maternity Care, a quick decision has to be made about the most suitable delivery type for the current patient. Guidelines are followed by physicians to support that decision; however, those practice recommendations are limited and underused. In the last years, caesarean delivery has been pursued in over 28% of pregnancies, and other operative techniques regarding specific problems have also been excessively employed. This study identifies obstetric and pregnancy factors that can be used to predict the most appropriate delivery technique, through the induction of data mining models using real data gathered in the perinatal and maternal care unit of Centro Hospitalar of Oporto (CHP). Predicting the type of birth envisions high-quality services, increased safety and effectiveness of specific practices to help guide maternity care decisions and facilitate optimal outcomes in mother and child. In this work was possible to acquire good results, achieving sensitivity and specificity values of 90.11% and 80.05%, respectively, providing the CHP with a model capable of correctly identify caesarean sections and vaginal deliveries.

Using data mining for prediction of hospital length of stay: an application of the CRISP-DM Methodology

Hospitals are nowadays collecting vast amounts of data related with patient records. All this data hold valuable knowledge that can be used to improve hospital decision making. Data mining techniques aim precisely at the extraction of useful knowledge from raw data. This work describes an implementation of a medical data mining project approach based on the CRISP-DM methodology. Recent real-world data, from 2000 to 2013, were collected from a Portuguese hospital and related with inpatient hospitalization. The goal was to predict generic hospital Length Of Stay based on indicators that are commonly available at the hospitalization process (e.g., gender, age, episode type, medical specialty). At the data preparation stage, the data were cleaned and variables were selected and transformed, leading to 14 inputs. Next, at the modeling stage, a regression approach was adopted, where six learning methods were compared: Average Prediction, Multiple Regression, Decision Tree, Artificial Neural Network ensemble, Support Vector Machine and Random Forest. The best learning model was obtained by the Random Forest method, which presents a high quality coefficient of determination value (0.81). This model was then opened by using a sensitivity analysis procedure that revealed three influential input attributes: the hospital episode type, the physical service where the patient is hospitalized and the associated medical specialty. Such extracted knowledge confirmed that the obtained predictive model is credible and with potential value for supporting decisions of hospital managers.

Tunnel engineering – influence of the type and the quantity of measurements in the back analysis of geomechanical parameters

The monitoring data collected during tunnel excavation can be used in inverse analysis procedures in order to identify more realistic geomechanical parameters that can increase the knowledge about the interested formations. These more realistic parameters can be used in real time to adapt the project to the real structure in situ behaviour. However, monitoring plans are normally designed for safety assessment and not especially for the purpose of inverse analysis. In fact, there is a lack of knowledge about what types and quantity of measurements are needed to succeed in identifying the parameters of interest. Also, the optimisation algorithm chosen for the identification procedure may be important for this matter. In this work, this problem is addressed using a theoretical case with which a thorough parametric study was carried out using two optimisation algorithms based on different calculation paradigms, namely a conventional gradient-based algorithm and an evolution strategy algorithm. Calculations were carried for different sets of parameters to identify several combinations of types and amount of monitoring data. The results clearly show the high importance of the available monitoring data and the chosen algorithm for the success rate of the inverse analysis process.

Search for type-III Seesaw heavy leptons in pp collisions at s√=8 TeV with the ATLAS Detector

A search for the pair-production of heavy leptons (N0,L±) predicted by the type-III seesaw theory formulated to explain the origin of small neutrino masses is presented. The decay channels N0→W±l∓ (ℓ=e,μ,τ) and L±→W±ν (ν=νe,νμ,ντ) are considered. The analysis is performed using the final state that contains two leptons (electrons or muons), two jets from a hadronically decaying W boson, and large missing transverse momentum. The data used in the measurement correspond to an integrated luminosity of 20.3fb−1 of pp collisions at s√=8 TeV collected by the ATLAS detector at the LHC. No evidence of heavy lepton pair-production is observed. Heavy leptons with masses below 325--540 GeV are excluded at the 95% confidence level, depending on the theoretical scenario considered.

Nonparametric estimation of the survival function for ordered multivariate failure time data: a comparative study

In longitudinal studies of disease, patients may experience several events through a follow-up period. In these studies, the sequentially ordered events are often of interest and lead to problems that have received much attention recently. Issues of interest include the estimation of bivariate survival, marginal distributions and the conditional distribution of gap times. In this work we consider the estimation of the survival function conditional to a previous event. Different nonparametric approaches will be considered for estimating these quantities, all based on the Kaplan-Meier estimator of the survival function. We explore the finite sample behavior of the estimators through simulations. The different methods proposed in this article are applied to a data set from a German Breast Cancer Study. The methods are used to obtain predictors for the conditional survival probabilities as well as to study the influence of recurrence in overall survival.

Spatial-temporal modellization of the NO2 concentration data through geostatistical tools

The nitrogen dioxide is a primary pollutant, regarded for the estimation of the air quality index, whose excessive presence may cause significant environmental and health problems. In the current work, we suggest characterizing the evolution of NO2 levels, by using geostatisti- cal approaches that deal with both the space and time coordinates. To develop our proposal, a first exploratory analysis was carried out on daily values of the target variable, daily measured in Portugal from 2004 to 2012, which led to identify three influential covariates (type of site, environment and month of measurement). In a second step, appropriate geostatistical tools were applied to model the trend and the space-time variability, thus enabling us to use the kriging techniques for prediction, without requiring data from a dense monitoring network. This method- ology has valuable applications, as it can provide accurate assessment of the nitrogen dioxide concentrations at sites where either data have been lost or there is no monitoring station nearby.

Local non-negative initial data scalar characterization of the Kerr solution

For any vacuum initial data set, we define a local, non-negative scalar quantity which vanishes at every point of the data hypersurface if and only if the data are Kerr initial data. Our scalar quantity only depends on the quantities used to construct the vacuum initial data set which are the Riemannian metric defined on the initial data hypersurface and a symmetric tensor which plays the role of the second fundamental form of the embedded initial data hypersurface. The dependency is algorithmic in the sense that given the initial data one can compute the scalar quantity by algebraic and differential manipulations, being thus suitable for an implementation in a numerical code. The scalar could also be useful in studies of the non-linear stability of the Kerr solution because it serves to measure the deviation of a vacuum initial data set from the Kerr initial data in a local and algorithmic way.

Big data e data analysis: visualização de informação

Dissertação de mestrado integrado em Engenharia e Gestão de Sistemas de Informação

Type 2 diabetes-related variants influence the risk of developing multiple myeloma: results from the IMMEnSE consortium

Type 2 diabetes (T2D) has been suggested to be a risk factor for multiple myeloma (MM), but the relationship between the two traits is still not well understood. The aims of this study were to evaluate whether 58 genome-wide-association-studies (GWAS)-identified common variants for T2D influence the risk of developing MM and to determine whether predictive models built with these variants might help to predict the disease risk. We conducted a case–control study including 1420 MM patients and 1858 controls ascertained through the International Multiple Myeloma (IMMEnSE) consortium. Subjects carrying the KCNQ1rs2237892T allele or the CDKN2A-2Brs2383208G/G, IGF1rs35767T/T and MADDrs7944584T/T genotypes had a significantly increased risk of MM (odds ratio (OR)=1.32–2.13) whereas those carrying the KCNJ11rs5215C, KCNJ11rs5219T and THADArs7578597C alleles or the FTOrs8050136A/A and LTArs1041981C/C genotypes showed a significantly decreased risk of developing the disease (OR=0.76–0.85). Interestingly, a prediction model including those T2D-related variants associated with the risk of MM showed a significantly improved discriminatory ability to predict the disease when compared to a model without genetic information (area under the curve (AUC)=0.645 vs AUC=0.629; P=4.05×10-06). A gender-stratified analysis also revealed a significant gender effect modification for ADAM30rs2641348 and NOTCH2rs10923931 variants (Pinteraction=0.001 and 0.0004, respectively). Men carrying the ADAM30rs2641348C and NOTCH2rs10923931T alleles had a significantly decreased risk of MM whereas an opposite but not significant effect was observed in women (ORM=0.71 and ORM=0.66 vs ORW=1.22 and ORW=1.15, respectively). These results suggest that TD2-related variants may influence the risk of developing MM and their genotyping might help to improve MM risk prediction models.